One install, three MCPs: graph chat memory + semantic code search + backend. Benchmarked with claude-haiku-4-5: -19% tokens & 23× faster on memory-recall questions.
python memory mcp pagerank knowledge-graph agents openie rag apache-2 pgvector graphrag model2vec hipporag claude-code semble insforge chat-memory
-
Updated
May 6, 2026 - Python